Diphenylacetylene-linked peptide strands induce bidirectional β-sheet formation

Article abstract of DOI:10.1002/anie.201309353

In the search for synthetic mimics of protein secondary structures relevant to the mediation of protein-protein interactions, we have synthesized a series of tetrasubstituted diphenylacetylenes that display β-sheet structures in two directions. Extensive

Full text of DOI:10.1002/anie.201309353

.
Angewandte
Communications
DOI: 10.1002/anie.201309353
Peptidomimetics
Hot Paper
Diphenylacetylene-Linked Peptide Strands Induce Bidirectional
b-Sheet Formation**
Hannah Lingard, Jeongmin T. Han, Amber L. Thompson, Ivanhoe K. H. Leung,
Richard T. W. Scott, Sam Thompson,* and Andrew D. Hamilton*
Abstract: In the search for synthetic mimics of protein
secondary structures relevant to the mediation of protein–
protein interactions, we have synthesized a series of tetrasub-
stituted diphenylacetylenes that display b-sheet structures in
two directions. Extensive X-ray crystallographic and NMR
solution phase studies are consistent with these proteomimetics
adopting sheet structures, displaying both hydrophobic and
hydrophilic amino acid side chains.
A
key element of the structure and function of b-sheet
domains in proteins is the almost perpendicular presentation
of side-chain residues in alternating periodicity above and
below the plane of the hydrogen-bonded network (Fig-
ure 1a). These residues are critical in forming intramolecular
contacts with other domains in many folded proteins, or in
determining the intermolecular associations that stabilize
certain protein–protein interactions (PPIs). b-Strand/b-strand
interactions[1] between different protein chains to form
interchain b-sheets account for around 16% of all PPIs and
have been implicated in a range of diseases, including
cancer.[2] Furthermore, human pathologies, such as Alzheim-
erꢀs and Parkinsonꢀs diseases, and dialysis-related amyloidosis
are associated with amyloid fiber formation involving pri-
marily the intermolecular stacking interaction of b-strands
and b-sheets.[3]
Figure 1. Strategies to template b-sheet formation.
There is much interest in the development of molecules
capable of mimicking the structure and recognition mecha-
nisms of b-sheets as potential disruptors of PPIs.[4] In addition,
the ability to construct easily variable functionalized surfaces,
in the manner of an extended b-sheet, has many potential
applications in supramolecular chemistry for the formation of
new materials or catalysts. Whereas much progress has been
made in the development of synthetic mimics of a-helices,[5]
the development of analogous b-sheet constructs has been
slower. In a seminal study, Kemp and co-workers described
a b-turn mimic derived from a substituted diphenylacetylene
that templates intramolecular hydrogen-bonded b-sheet for-
mation.[6] This broad approach of covalently linking one end
of a b-sheet has been extended by others using tolan
derivatives,[7] substituted biphenyl,[8] dibenzofuran,[9] oligoan-
thranilamide,[10] and urea scaffolds,[11] as well as more unusual
motifs such as ferrocene[12] or other metal-chelating groups[13]
to template the two strands. Others have exploited the simple
strategy of cyclizing peptide or peptidomimetic strands[14] to
form a length of b-sheet flanked by two b-turn units (Fig-
ure 1b). Unnatural amino acids have been used to induce b-
sheet formation (Figure 1c)[15] and cross-linking strands using
disulfide bonds has also been employed (Figure 1d).[16]
In the course of our investigations of synthetic analogues
of protein secondary structure, we have reported the use of
completely unnatural backbone oligomers as b-strand mim-
etics.[17] Herein, we report a different strategy, using a tetra-
substituted diphenylacetylene as both a rigidifying and a link-
ing spacer between two peptide strands. This approach has the
potential to promote sheet formation in two directions, and
differs structurally from many existing templates that use
terminal linkers. Incorporating amino acids into each quad-
rant results in a scaffold from which four substituents project
from a single face in conceptual analogy to the positioning of
residues on a b-sheet (Figure 1e).
[*] Dr. H. Lingard, J. T. Han, Dr. A. L. Thompson, Dr. I. K. H. Leung,
Dr. R. T. W. Scott, Dr. S. Thompson, Prof. A. D. Hamilton
Chemistry Research Laboratory, University of Oxford
12 Mansfield Road, Oxford, OX1 3TA (UK)
E-mail: sam.thompson@chem.ox.ac.uk
andrew.hamilton@chem.ox.ac.uk
Homepage: http://hamilton.chem.ox.ac.uk
[**] We thank The University of Oxford (HKL, ST) for funding, Dr.
Martin D. Smith and Dr. Peter C. Knipe for helpful discussions, and
Diamond Light Source for an award of beamtime (MT7768).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201309353.
3650
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 3650 –3653

Angewandte
Chemie
As alanine is a good example of an amino acid preferen-
tially adopting the antiparallel b-sheet structure,[18] we used it
as a model residue in a convergent synthesis that readily
allows the incorporation of other amino acids. Reduction of
nitro benzoate 1[19] provided aniline 2, which was coupled with
N-Boc-alanine 3 (Boc = tert-butoxycarbonyl) to give iodide 4
(not shown, see the Supporting Information). Iodide 4 is
a common precursor for both fragments of the target
molecule. A Sonogashira reaction with TMS-acetylene fol-
lowed by basic removal of the silyl group gave the second
coupling fragment, alkyne 6. A second Sonogashira reaction
is used to link the fragments in reasonable yield. The structure
of two-residue mimic 7 was solved from single crystal X-ray
diffraction data,[20] and presented a broadly planar arrange-
ment with hydrogen bonds between the amide NH and the
=
ester carbonyl on each side of the diphenylacetylene. The C
À
O···N distances are between 3.10 and 2.84 ꢁ, with N H···O
distances of 2.15 and 2.04 ꢁ (dashed black lines in Scheme 1;
some hydrogens omitted for clarity). Installation of two
further amino acids began with basic hydrolysis of the methyl
esters of 7 to give a dicarboxylic acid followed by coupling
with amino amide 8, derived from l-alanine, to give fully
elaborated b-sheet analogue 9 (Scheme 1).
Figure 2. Solution-phase analysis in CDCl3. a) 1H shifts of Ca-hydro-
gens in sheet and open structures. ROESY experiment of 9 with
correlations shown in red (symmetry-related correlations ommitted).
Rate of [D4]methanol exchange for NH groups: b) sheet mimic 9,
c) control 10; Ha/Ha’ (blue), Hb/Hb’ (red), Hc/Hc’ (yellow), Hd/Hd’
(green).
those exposed to solvent.[22] Figure 2b shows the rates at
which amide hydrogens are exchanged for deuterons after the
addition of a small amount of [D4]methanol to a solution of 9
in CDCl3. Hd (terminal amide, green) and Hb (anilide, red)
exchange more slowly with the solvent than Ha (carbamate,
blue) and Hc (amide, yellow), which is consistent with the
hydrogen-bonding network proposed. Figure 2c shows the
rates of exchange of the four equivalent hydrogens in model
system 10, which is incapable of intramolecular hydrogen
bonding. In this case, hydrogens Hd’ and Hb’ (green and red)
exchange much faster than the equivalent hydrogens in sheet
analogue 9, thus supporting the fact that these hydrogens are
shielded from the solvent in the latter.
Although it was not possible to obtain satisfactory single
crystals of 9, substitution of the Boc groups with acetyl groups
gave crystals of 11 suitable for X-ray diffraction experiments
(Figure 3).[20] Consistent with the solution-phase studies, four
intramolecular hydrogen bonds stabilize the desired b-sheet
conformation, with the alanine side-chains displayed on the
Scheme 1. Synthesis of a four-alanine residue sheet mimic, with an X-
ray diffraction structure of precursor 7.
=
A ROESY spectrum of 9 in chloroform (Figure 2a)
showed correlations between the Ca hydrogens of the two
alanine groups on the linked strands of the b-sheet, as well as
between the tert-butyl group hydrogens of the Boc group and
the iso-butyl hydrogens of the amide. These indicate that the
ends of the two strands are in close proximity, which is
consistent with the iso-butyl amide NH forming a hydrogen
bond to the Boc carbonyl on the adjacent strand. A
same face of the molecule. The C O···N distances range
À
between 3.03 and 2.75 ꢁ, with N H···O distances between
2.21 and 1.90 ꢁ. The interstrand Ca/Ca distances of 4.0 and
4.5 ꢁ are within the range of those found in natural b-
sheets.[23] Additional intermolecular NH···O C hydrogen
=
bonds were formed with adjacent molecules in the crystal,
resulting in supramolecular structures featuring unidirec-
tional side-chain projection.
À
comparison of the Ca H shifts of iodide 10 with those of
To demonstrate the generality of this approach for the
incorporation of natural or unnatural amino acids, we
followed an analogous synthetic route to prepare the four-
residue mimic 12, incorporating two lysine and two glutamic
acid residues (Figure 4; see the Supporting Information for
the synthesis of these compounds).
sheet mimic 9 shows that both have moved downfield, from
4.6 to 5.2 ppm and 4.4 to 5.5 ppm, respectively, which is
consistent with the two strands of 9 having formed a b-sheet
(Figure 2).[21]
Further support for this structure comes from deuterium
exchange experiments where the hydrogen-bonded amide-
NH groups would be expected to exchange less rapidly than
Hydrophilic mimetic 12 was insoluble in chloroform, and
so we first probed its solution-phase conformation by
Angew. Chem. Int. Ed. 2014, 53, 3650 –3653
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3651

.
Angewandte
Communications
“protected” within intramolecular H-bonds. The addition of
2,2,2-trifluoroethanol to 12 in [D6]DMSO caused Hb and Hd
to move slightly downfield, whereas Ha and Hc moved upfield
to a greater degree. These observations can be explained by
2,2,2-trifluoroethanol H-bonding to the solvent-exposed car-
bonyl groups of amides bearing Hb and Hd, thus leading to
electron withdrawal and a downfield shift. The upfield shifts
of Ha and Hc likely come from a reduction in their H-bonding
to [D6]DMSO as a result of competition with added 2,2,2-
trifluoroethanol.[27]
To explore the conformational behavior of sheet mimic 12
in aqueous media, we added H2O incrementally to
a [D6]DMSO solution and recorded ROESY spectra, and
temperature-dependent coefficients of the backbone amide
NHs. We observed the same interstrand correlations as those
in [D6]DMSO, at up to 40% H2O. Increasing the ratio of H2O
to 70% required a high level of solvent suppression that led to
silencing of the region in which the key resonances of the Ca-,
and benzylic-hydrogens reside, making it difficult to gather
further data. However, over this range (0–70% H2O), the
temperature-dependent coefficients of Hb and Hd did not
change significantly,[28] which suggests the maintenance of
interstrand hydrogen bonding[29] and thus the persistence of
a sheet conformation (see the Supporting Information,
Chapter 4 “Conformational analysis”).
Figure 3. a) Sheet mimic 11; b) top and, c) side-projections of X-ray
structure; some hydrogens omitted for clarity, alanine side chains
highlighted in orange.
In conclusion, we have developed a general strategy to
protein secondary structure mimics capable of displaying b-
sheet-like structures in two directions, and have demonstrated
its efficiency for the incorporation of both hydrophobic and
hydrophilic amino acid side chains. The use of a central
rigidifying and templating group covalently attached to four
peptide strands is distinct from extant approaches to b-sheet
mimicry, and thus offers a new method for the display of
protein surfaces. Solid- and solution-phase conformational
studies in a range of solvent systems are consistent with the
interstrand hydrogen bonds between peptide chains, and the
projection of amino acid side chains, such as those found in
natural b-sheets. Development of this class of b-sheet mimic
for use in mediating PPIs and as a basis for forming analogues
of higher order protein structure is underway.
Figure 4. Solution-phase analysis of mimic 12 and control compound
13. [D6]DMSO ROESY correlations shown as red arrows (dashed
arrows represent weaker interactions). For color coding of amide
hydrogens for variable temperature and titration studies, see text.
conducting a series of NMR experiments in [D6]DMSO and
compared the results to those of control compound 13, which
is incapable of intramolecular H-bonding (Figure 4; full data
and analysis are given in the Supporting Information). A
ROESY spectrum of 12 showed cross-strand correlations
consistent with the population of a sheet conformation
analogous to those observed for 9. Further evidence for the
existence of a b-sheet structure in solution was provided by
a variable temperature NMR study. Temperature-dependent
coefficients (Dd/DK) for Hb (anilide, red) and Hd (terminal
amide, green) of À0.89 and À3.85 ppbKÀ1, respectively, are
consistent with intramolecular H-bonding, whereas coeffi-
Received: October 26, 2013
Revised: December 17, 2013
Published online: February 19, 2014
Keywords: beta-sheets · peptidomimetics · protein surfaces ·
cients of À7.45 and À6.88 ppbKÀ1 for H (carbamate, blue)
.
a
protein-protein interactions · secondary structures
and Hc (amide, yellow), respectively, suggest interaction with
solvent only.[24]
The presence of a H-bonding network was also probed by
the addition of both CDCl3 and 2,2,2-trifluoroethanol to
[D6]DMSO solutions. Upon increasing titration with CDCl3,
Hb is essentially unperturbed, whereas the upfield shift of Hd
is markedly smaller than those of Ha and Hc.[25] This behavior
is consistent with previous studies in which a lowering of the
proportion of a powerfully H-bond-accepting solvent leads to
greater perturbations for solvent-exposed amide NHs relative
to those in intramolecular H-bonds.[26] Under these conditions
it is expected that solvent-exposed NH groups will be more
susceptible to a change in solvent composition than those
[1] H. Yin, G. Lee, H. S. Park, G. A. Payne, J. M. Rodriguez, S. M.
Sebti, A. D. Hamilton, Angew. Chem. 2005, 117, 2764 – 2767;
Angew. Chem. Int. Ed. 2005, 44, 2704 – 2707.
[2] J. S. Nowick, D. M. Chung, K. Maitra, S. Maitra, K. D. Stigers, Y.
Sun, J. Am. Chem. Soc. 2000, 122, 7654 – 7661.
[3] F. Chiti, C. M. Dobson, Nat. Chem. Biol. 2009, 5, 15 – 22.
[4] W. A. Loughlin, J. D. A. Tyndall, M. P. Glenn, T. A. Hill, D. P.
Fairlie, Chem. Rev. 2010, 110, PR32 – PR69.
[5] a) M. K. P. Jayatunga, S. Thompson, A. D. Hamilton, Bioorg.
Med. Chem. Lett. 2014, 24, 717 – 724; b) V. Azzarito, K. Long,
N. S. Murphy, A. J. Wilson, Nat. Chem. 2013, 5, 161 – 173; c) S.
Thompson, R. Vallinayagam, M. J. Adler, R. T. W. Scott, A. D.
3652
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 3650 –3653

Angewandte
Chemie
Hamilton, Tetrahedron 2012, 68, 4501 – 4505; d) S. Thompson,
A. D. Hamilton, Org. Biomol. Chem. 2012, 10, 5780 – 5782.
[6] a) D. S. Kemp, Z. Q. Li, Tetrahedron Lett. 1995, 36, 4175 – 4178;
b) D. S. Kemp, Z. Q. Li, Tetrahedron Lett. 1995, 36, 4179 – 4180.
[7] a) A. C. Spivey, J. McKendrick, R. Srikaran, B. A. Helm, J. Org.
Chem. 2003, 68, 1843 – 1851; b) D. A. Offermann, J. E. McKen-
drick, J. J. P. Sejberg, B. Mo, M. D. Holdom, B. A. Helm, R. J.
Leatherbarrow, A. J. Beavil, B. J. Sutton, A. C. Spivey, J. Org.
Chem. 2012, 77, 3197 – 3214.
[8] C. L. Nesloney, J. W. Kelly, J. Am. Chem. Soc. 1996, 118, 5836 –
5845.
[9] H. A. Lashuel, S. R. LaBrenz, L. Woo, L. C. Serpell, J. W. Kelly,
J. Am. Chem. Soc. 2000, 122, 5262 – 5277.
Thompson, R. T. W. Scott, A. D. Hamilton, Chem. Commun.
2012, 48, 9834 – 9836.
[18] B. Rost, C. Sander, J. Mol. Biol. 1993, 232, 584 – 599.
[19] E. C. Y. Woon, A. Dhami, M. F. Mahon, M. D. Threadgill,
Tetrahedron 2006, 62, 4829 – 4837.
[20] CCDC 915930 (7) & 915931 (11) contain the supplementary
crystallographic data for this paper. These data can be obtained
free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
[21] D. S. Wishart, B. D. Sykes in Methods in Enzymology (Ed.: T. L.
James), Academic Press, New York, 1994, 363.
[22] L. R. Steffel, T. J. Cashman, M. H. Reutershan, B. R. Linton, J.
Am. Chem. Soc. 2007, 129, 12956 – 12957.
[23] C.-I. Brꢂndꢃn, J. Tooze, Introduction to Protein Structure,
Garland Publishing, New York, 1999.
[10] Y. Hamuro, A. D. Hamilton, Bioorg. Med. Chem. 2001, 9, 2355 –
2363.
[24] a) E. S. Stevens, N. Sugawara, G. M. Bonora, C. Toniolo, J. Am.
Chem. Soc. 1980, 102, 7048 – 7050; b) A. Ballardin, A. J. Fisch-
man, W. A. Gibbons, J. Roy, I. L. Schwartz, C. W. Smith, R.
Walter, H. R. Wyssbrod, Biochemistry 1978, 17, 4443 – 4454;
c) Coefficients for Ha’-Hd’ in control molecule 13 were between
À4.92 and À7.33 ppbKÀ1, suggesting interactions with solvent
only.
[11] J. S. Nowick, E. M. Smith, G. Noronha, J. Org. Chem. 1995, 60,
7386 – 7387.
[12] S. Chowdhury, G. Schatte, H.-B. Kraatz, Angew. Chem. 2008,
120, 7164 – 7167; Angew. Chem. Int. Ed. 2008, 47, 7056 – 7059.
[13] A. C. Laungani, J. M. Slattery, I. Krossing, B. Breit, Chem. Eur. J.
2008, 14, 4488 – 4502.
[14] a) J. A. Robinson, Acc. Chem. Res. 2008, 41, 1278; b) J. A.
Robinson, S. DeMarco, F. Gombert, K. Moehle, D. Obrecht,
Drug Discovery Today 2008, 13, 944 – 951; c) F. Freire, S. H.
Gellman, J. Am. Chem. Soc. 2009, 131, 7970 – 7972; d) F. Freire,
S. H. Gellman, J. Am. Chem. Soc. 2011, 133, 12318.
[15] J. S. Nowick, Acc. Chem. Res. 2008, 41, 1319 – 1330.
[16] A. M. Almeida, R. Li, S. H. Gellman, J. Am. Chem. Soc. 2012,
134, 75 – 78.
[25] Increasing CDCl3 content in DMSO from 44 to 90% gave the
following perturbations: Ha À24.3, Hb + 0.5, Hc À16.7, Hd
À3.9 ppbKÀ1
.
[26] H. Kessler, Angew. Chem. 1982, 94, 509 – 520; Angew. Chem. Int.
Ed. Engl. 1982, 21, 512 – 523.
[27] a) M. Llinas, M. P. Klein, J. Am. Chem. Soc. 1975, 97, 4731 –
4737; b) D. W. Urry, M. M. Long, CRC Crit. Rev. Biochem. 1976,
4, 1 – 45.
[17] a) P. N. Wyrembak, A. D. Hamilton, J. Am. Chem. Soc. 2009,
131, 4566 – 4567; b) A. G. Jamieson, D. Russell, A. D. Hamilton,
Chem. Commun. 2012, 48, 3709 – 3711; c) C. L. Sutherell, S.
[28] On going from 0 to 70% H2O, Hb and Hd moved upfield 1.3 and
0.95 ppbKÀ1, respectively (to À1.5 and À4.8 ppbKÀ1).
[29] T. Cierpicki, J. Otlewski, J. Biomol. NMR 2001, 21, 249 – 261.
Angew. Chem. Int. Ed. 2014, 53, 3650 –3653
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3653